The brief

Analyst view

• Series C August 2026: Nvidia invests $1B and licenses Model Factory for $6B at $12B valuation.
• Laguna S 2.1, released July 21, 2026, achieves 78.5% on SWE-Bench Multilingual, outperforming models 10–20x its size.
• The model activates 8 billion of 118 billion parameters per token and fits on a single NVIDIA DGX Spark when quantized.

Technical approach

As reported

Sparse MoE

Mixture of Experts for efficiency

Laguna S 2.1 uses 256 routed experts with a 1 million+ token context window, activating only 8B of 118B parameters per token for inference speed and cost.

Coding focus

Built for software engineering

Models trained on code with FP8 precision RL. XS runs locally; M.1 targets enterprise.

Open weights

Permissive licensing

Laguna S 2.1 released under OpenMDW-1.1 license; available via Ollama, vLLM, and Hugging Face, with free preview access via OpenRouter.